document how auto reload of JCasC changes works about helm-charts HOT 5 OPEN

This is a good issue to address this in documentation clearly.

My use case is to install jenkins via helm chart and keep it running like before, and want to configure them manually via GUI, so we disable the sidecar for autoupdate and reuse the pvc like below

persistence:
    enabled: true
    storageClass: nfs
    existingClaim: jenkins-nfs-permament
master:
    sidecars:
        configAutoReload:
          enabled: false
# ...
    JCasC:
        enabled: true
        configScripts:
          jenkins: |
            jenkins:
              systemMessage: |
                Welcome to our Jenkins CI server.
          views: |
              jenkins:
                  views:
                    - list:
                        columns:
                          - "status"
                          - "jobName"
                          - "lastSuccess"
                          - "lastFailure"
                          - "lastDuration"
                        includeRegex: "^test_.*"
                        name: "Sandbox"

And we noticed after restart the pod (simulate the crash of jenkins pod), some configurations are reloaded while some are kept, it will be good to let us know the strategy and way to check.

So far it feels if it defines in $CASC_JENKINS_CONFIG, it will be reloaded like jobs and system configuration.

But some parts are tricky to understand like

• if I define some job views in values.yaml (see above), I can't manually to add new views in jenkins, they will be reset always after restart.

And it is good to see new generated jobs are kept (no matter they are from seed job or manually added jobs)

Therefore in this issue, hopefully we can cover those things in documentation.

from helm-charts.

Hi @torstenwalter Can you assign this to me?

from helm-charts.

I just ran into this not working as well. I was about to open a feature request around this before I saw this one.

Perhaps in a future release, we could do away with the sidecar (which doesn't appear to be working properly at the moment) and use pod annotations (separate from .Values.controller.podAnnotations) based on the md5sum of each .Values.controller.JCasC.configScripts instead.

The only downside I can think of to this approach is that JCasC changes will require a couple of minutes of Jenkins downtime (which is what the sidecar is trying to avoid).

from helm-charts.

I confirm that sidecars is also not working for me, seems to be completely broken.

On the wiki, we're advised to enable rbac, but it still fails to register the values.yaml changes.

So far, the only workaround that I have found is to "manually" or through CI/CD restart the pods, which causes a downtime overhead as @unacceptable has mentionned.

from helm-charts.

any update on this. seems to still be an issue.

from helm-charts.